It was theorized in the 1960's that tumor cells bear specific antigens (TSA) which are not present on normal cells and that the immune response to these antigens might enable an individual to reject a tumor. It was later suggested that the immune response to TSA could be increased by introducing new immunological determinants on cells. Mitchison, Transplant. Proc., 1970, 2, 92. Such a "helper determinant", which can be a hapten, a protein, a viral coat antigen, a transplantation antigen, or a xenogenous cell antigen, could be introduced into a population of tumor cells. The cells would then be injected into an individual who would be expected to be tolerant to the growth of unmodified tumor cells. Clinically, the hope was that an immunologic reaction would occur against the helper determinants, as a consequence of which the reaction to the accompanying TSA is increased, and tumor cells which would otherwise be tolerated are destroyed. Mitchison, supra, also suggests several modes of action of the helper determinants including 1) that the unmodified cells are merely attenuated, in the sense that their growth rate is slowed down or their susceptibility to uinrnunologic attack increased; 2) that helper determinants merely provide points of attack and so enable the modified cells to be killed by an immune response not directed against TSA; 3) that the helper determinants have an adjuvant action such as binding to an antibody or promoting localization of the cells in the right part of the body for immunization, in particular, in lymph nodes.
Fujiwara et al., J. Immunol., 1984a, 132, 1571 showed that certain haptenized tumor cells, i.e., tumor cells conjugated with the hapten trinitrophenyl (TNP), could induce systemic immunity against unmodified tumor cells in a murine system, provided that the mice were first sensitized to the hapten in the absence of hapten-specific suppressor T cells. Spleen cells from the treated mice completely and specifically prevented the growth of tumors in untreated recipient animals. Flood et al., J. Immunol., 1987, 138, 3573 showed that mice immunized with a TNP-conjugated, ultraviolet light-induced "regressor" tumor were able to reject a TNP-conjugated "progressor" tumor that was otherwise non-immunologic. Moreover, these mice were subsequently resistant to challenge with unconjugated "progressor" tumor. In another experimental system, Fujiwara et al., J. Immunol., 1984b, 133, 510 demonstrated that mice sensitized to trinitrochlorobenzene (TNCB) after cyclophosphamide pretreatment could be cured of large (10 mm) tumors by in situ haptenization of tumor cells; subsequently, these animals were specifically resistant to challenge with unconjugated tumor cells.
The teachings of Fujiwara et al. differ from the present invention for several reasons including the following: A. The cells used in Fujiwara's composition were derived from induced transplantable murine tumors--not from spontaneous human tumors; B. Fujiwara's composition is used in immunoprophylaxis--the present invention uses immunotherapy; C. Fujiwara's composition is administered as a local therapy--the present invention is administered by systemic inoculation; and D. Fujiwara's composition did not result in tumor regression--the composition of the present invention results in regression and/or prolonged survival for at least a substantial portion of the patients treated.
The existence of T cells which cross-react with unmodified tissues has recently been demonstrated. Weltzien and coworkers have shown that class I MHC-restricted T cell clones generated from mice immunized with TNP-modified syngeneic lymphocytes respond to MHC-associated, TNP-modified "self" peptides. Ortmann, B., et al., J. Immunol., 1992, 148, 1445. In addition, it has been established that immunization of mice with TNP-modified lymphocytes results in the development of splenic T cells that exhibit secondary proliferative and cytotoxic responses to TNP-modified cells in vitro. Shearer, G. M. Eur. J. Immunol., 1974, 4, 527. The potential of lymphocytes elicited by immunization with DNP- or TNP-modified autologous cells to respond to unmodified autologous cells is of considerable interest because it may be relevant to two clinical problems: 1) drug-induced autoimmune disease, and 2) cancer immunotherapy. In regard to the former, it has been suggested that ingested drugs act as haptens, which combine with normal tissue protein forming immunogenic complexes that are recognized by T cells.
Tsutsui, H., et al., J. Immunol., 1992, 149, 706. Subsequently, autoimmune disease, e.g., systemic lupus erythematosus, can develop and continue even after withdrawal of absence of the offending drug. This would imply the eventual generation of T lymphocytes that cross-react with unmodified tissues.
The common denominator of these experiments is sensitization with hapten in a milieu in which suppressor cells are not induced. Spleen cells from cyclophosphamide pretreated, TNCB-sensitized mice exhibited radioresistant "amplified helper function" i.e., they specifically augmented the in vitro generation of anti-TNP cytotoxicity. Moreover, once these amplified helpers had been activated by in vitro exposure to TNP-conjugated autologous lymphocytes, they were able to augment cytotoxicity to unrelated antigens as well, including tumor antigens (Fujiwara et al., 1984b). Flood et al., (1987), supra, showed that this amplified helper activity was mediated by T cells with the phenotype Lyt-1.sup.+, Lyt.sup.- 2.sup.-, L3T4.sup.+, I.sup.- J.sup.+ and suggests that these cells were contrasuppressor cells, a new class of immunoregulatory T cell.
Immunotherapy of patients with melanoma had shown that administration of cyclophosphamide, at high dose (1000 mg/M.sup.2) or low dose (300 mg/M.sup.2), three days before sensitization with the primary antigen keyhole limpet hemocyanin markedly augments the acquisition of delayed type hypersensitivity to that antigen (Berd et al., Cancer Res., 1982, 42, 4862; Cancer Res., 1984a, 44, 1275). Low dose cyclophosphamide pretreatment allows patients with metastatic melanoma to develop delayed type hypersensitivity to autologous melanoma cells in response to injection with autologous melanoma vaccine (Berd et al., Cancer Res., 1986, 46, 2572). The combination of low dose cyclophosphamide and vaccine can produce clinically important regression of metastatic tumor (Berd et al. (1986), supra; Cancer Invest., 1988a, 6, 335). Cyclophosphamide administration results in reduction of peripheral blood lymphocyte non-specific T suppressor function (Berd et al., Cancer Res., 1984b, 44, 5439; Cancer Res., 1987, 47, 3317), possibly by depleting CD4+, CD45R+ suppressor inducer T cells (Berd et al., Cancer Res., 1988b, 48, 1671). The antitumor effects of this immunotherapy regimen appear to be limited by the excessively long interval between the initiation of vaccine administration and the development of delayed type hypersensitivity to the tumor cells (Berd et al., Proc. Amer. Assoc. Cancer Res., 1988c, 29, 408 (#1626)). Therefore, there remained a need to increase the therapeutic efficiency of such a vaccine to make it more immunogenic.
Most tumor immunologists now agree that T lymphocytes, white cells responsible for tumor immunity, infiltration into the tumor mass is a prerequisite for tumor destruction by the immune system. Consequently, a good deal of attention has been focused on what has become known as "TIL" therapy, as pioneered by Dr. Stephen Rosenberg at NCL Dr. Rosenberg and others have extracted from human cancer metastases the few T lymphocytes that are naturally present and greatly expanded their numbers by culturing them in vitro with Interleukin 2 (IL2), a growth factor for T lymphocytes. Topalian et al., J. Clin. Oncol., 1988, 6, 839. However this therapy has not been very effective because the injected T cells are limited in their ability to "home" to the tumor site.
The ability of high concentrations of IL2 to induce lymphocytes to become non-specifically cytotoxic killer cells has been exploited therapeutically in a number of studies (Lotze et al., J. Biol. Response, 1982, 3, 475; West et al., New Engl. J. Med., 1987, 316, 898). However, this approach has been limited by the severe toxicity of high dose intravenous IL2. Less attention has been given to the observation that much lower concentrations of IL2 can act as an immunological adjuvant by inducing the expansion of antigen-activated T cells (Talmadge et al., Cancer Res., 1987, 47, 5725; Meuer et al., Lancet, 1989, 1, 15). Therefore, there remains a need to understand and attempt to exploit the use of IL2 as an immunological adjuvant.
Human melanomas are believed to express unique surface antigens recognizable by T lymphocytes. Old, L. J., Cancer Res., 1981, 41, 361; Van der Bruggen, P., etal., Science, 1991, 254, 1643; Mukhedi, B., etal., J. Immunol., 1986,136, 1888; and Anichini, A., et al., J. Immunol., 1989, 142, 3692. However, immunotherapeutic approaches prior to work done by the present inventor had been limited by the difficulty of inducing an effective T cell-mediated response to such antigens in vivo.
The present inventor obtained results including substantial tumor remission and prolongation of survival time with a haptenized rumor cell vaccine administered to patients with malignant melanoma.
There are several models proposed to explain what appears to be tolerance to human tumor-associated antigens. They include:
1) Tumor antigen-specific suppressor cells that down-regulated incipient anti-tumor responses. Mukhedi, et al., supra; Berendt, M. J. and R. J. North., J. Exp. Med., 1980, 151, 69.
2) Failure of human tumor cells to elicit T helper cells or to provide costimulatory signals to those T cells. Fearon, E. R., et al., Cell, 1990, 60, 397; Townsend, S. E. and J. P. Allison, Science, 1993, 259, 368; and
3) Reduced surface expression of major histocompatibility products on tumor cells which limits their recognition by T cells. Ruiter, D. J., Seminars in Cancer Biology, 1991, 2, 35. None of these hypotheses has yet been corroborated in a clinical system.
Regardless of whether such explanations are true or not, there is a continuing need for more effective treatment of various malignancies.
In regard to acute myelogenous leukemia (AML), the treatment for AML is divided into one or two initial induction phases and several courses of postremission, also known as consolidation, chemotherapy. Initial induction chemotherapy may induce a complete response in 55 to 88% of the patients, depending on the protocol used. However, the vast majority of these patients relapse, and the long-term (5 year+) survival of AML patients is only 20-30%. The addition of high-dose chemotherapy and bone marrow transplantation (BMT) to this therapeutic regime during the first remission can bring about modest improvements in result. For example, patients undergoing allogeneic BMT are afforded a 5 to 10% increase in the 5 year survival. However, the strict eligibility criteria for BMT (e.g., age, availability of an HLA-matched donor) severely limit the number of patients who can be treated. Once AML patients relapse, there is only a 30% chance of achieving a second remission, and very few of these patients remain disease-free in the long run. Treatment modalities on relapse include similar protocols to those used in achieving the first remission (induction therapy followed by several courses of consolidation chemotherapy), although high dose of a single agent and BMT can also be used (Keating et al.).
Experience with bone marrow transplantation has suggested that immunological rejection may play a role in the control of the disease. Graft-versus-host disease (GVHD) and relapse are the two main causes of death of patients treated with BMT.
The risk of relapse decreases if mild GVHD occurs (Horowitz et al.). Therefore it has been hypothesized that grafted lymphocytes are able to immunologically reject host leukemia cells (graft-versus-leukemia reaction, GVL). This GVL reaction could be mediated by a T-cell response against specific leukemia cell antigens, although immunogenic human leukemia antigens have not yet been demonstrated (the same is true for melanoma). It is known that human AML cells strongly express both class I and class II major histocompatibility complex (MHC) antigens (Ashman et al.; Andreasen et al.) which are prerequisites for the induction of CD8- and CD4-mediated T cell responses, respectively. However, induction of a T cell response targeted to leukemia cells has not been successful.
Several immunological approaches have been used for the treatment of acute leukemia (Foon et al.; Caron and Scheinberg). These approaches are divided into non-specific, such as Bacillus Calmene Guerin (BCG), interleukin-2, levamisole, methanol-extraction residue of tubercle bacillus, and specific, such as monoclonal antibodies and vaccines (harvested leukemia cells, cell free extracts and cultured cells). The majority of these studies have been performed in patients already in remission, in which immunotherapy would have to be successful in controlling residual disease.
In the late 1960's and early 1970's the research group R. Powles at St. Barthlomew's Hospital in England conducted a series of studies of vaccine treatment of AML patients after chemotherapy-induced remission (Powles, 1974; Powles et al., 1977). They used allogeneic AML cells with BCG as an adjuvant. Several trials were performed, all with small sample sizes (N=10-15). There was some prolongation of survival with chemotherapy+immunotherapy compared with chemotherapy alone, but no prolongation of relapse-free survival. No serious toxicity was observed; autoimmunity (e.g., toxicity to normal bone marrow) was not seen. In retrospect, there were a number of technical problems with these trials: 1) allogeneic, rather than autologous, leukemia cells were used; 2) the dose of leukemia cells in the vaccine was excessive (up to 10.sup.9 cells/dose); 3) the BCG dose was very high and BCG administration was separated by time and location from the leukemia cell vaccine; and 4) the vaccine was administered while the patients were receiving cytotoxic drugs (maintenance or consolidation chemotherapy).
The immunochemical basis of this phenomenon remains speculative, but several hypotheses are being tested. Kim and Jang (1992) have suggested that the lack of T cell response to a particular epitope may not be due to absence of a T cell repertoire, but rather to difficulty in generating the particular epitope. Martin et al. (1993) have explained their results by hypothesizing the existence of autoreactive T cells that escape thymic selection because of low affinity for "self" peptides. Hapten modification of such peptides may convert subdominant peptide epitopes into dominant determinants and thereby activate those T cells. Alternatively, hapten modification may facilitate antigen processing to generate the epitope.
This therapeutic regime results in elicitation of 1) T lymphocytes to infiltrate the tumor, 2) an inflammatory immune response to a tumor, and 3) a delayed-type hypersensitivity response to the tumor, and, ultimately, in at least a portion of the patient population in tumor regression (reduction of tumor burden).
Conventional attempts to treat human cancer have been unsuccessful. Administration of compositions, exemplified by those set forth above, failed to reliably induce the development of cell-mediated immunity as indicated by delayed-type hypersensitivity (DTH), T cell infiltration, and inflammatory immune response.
Accordingly, despite the number of theories proposed for the immunological effects reported in the treatments of cancer, there remains a need for a composition which, upon administration to an animal, is capable of eliciting T lymphocytes that infiltrate a tumor, eliciting an inflammatory immune response to a tumor, and eliciting a delayed-type hypersensitivity response to a tumor.